There is currently no consistently effective pharmacologic treatment for cocaine abuse. The proposed program of research is specifically designed to address this issue. We propose to conduct a series of medication trials of drug effectiveness with the following conceptual and methodological features: 1. Evaluation of drug effectiveness and measurement of outcomes specific to the stage of addiction treatment-early stabilization, rehabilitation, and aftercare; 2. Conduct of these medication trials in the context of a standardized and widely practiced form of psychosocial treatment - Cognitive Behavioral Relapse Prevention (CBRP) - such that the results will be relevant to the delivery of cocaine dependence treatment in many "real world" settings; 3. Inclusion of samples large and diverse enough to provide power to detect effects with relevant subgroups such as women and minorities. Other relevant subgroups (depending upon the particular medication) will include cocaine dependent and cocaine abuser patients (by DSM-IV criteria), those with and without Axis I disorders such as depression and anxiety, as well alcohol dependence; 4. consistent use of multiple measures of outcome including quantitative analysis of cocaine metabolites several times weekly via urinalysis; cocaine craving measured both by standard analog ratings and by a stimulated craving session under laboratory conditions yielding both physiological and self report data; and administration of the Clinician's Global Impression (CGI); and 5. participation in multi-clinic collaborative trials according to a common protocol when this is deemed appropriate by NIDA Medications Development Division (MDD). All studies will use a double blind placebo controlled design and include a minimum sample size of 100 evaluable subjects; half of the subjects will be women. Thus, it should be possible to independently ascertain outcomes for women as well as for men. State of the art assessment instruments including the Addiction Severity Index and the Treatment Services Review, developed in our Division of Addiction Studies, the Beck Depression Inventory, the Hamilton Depression scale, etc. will be employed. Methodological details of the initial two year study are described.